The "doctor-patient-computer triangle" has become a defining, and often isolating, feature of modern medicine. As physicians shift their gaze from the person across the exam table to the glowing screen of an electronic health record (EHR), the nuance of human connection is frequently lost to the demands of data entry. While administrative record-keeping is a non-negotiable pillar of clinical continuity, billing, and legal protection, the current reality suggests that doctors spend nearly twice as much time managing these digital records as they do face-to-face with their patients.
In Rhode Island, a significant shift is underway as hundreds of physicians begin integrating artificial intelligence to dismantle this triangle. By tasking AI with the heavy lifting of summarizing visits and drafting clinical notes, health systems are attempting to reclaim the time lost to the keyboard. This is not merely an upgrade in software; it is a fundamental pivot toward allowing clinicians to function as healers rather than administrative clerks.
Tomas Gregorio, the chief digital information officer at Care New England, recalls a time two decades ago when he first introduced tablets to assist with dictation. That early iteration successfully cut charting time from three hours a day to one-and-a-half. Today, the technology integrated into systems like Epic offers capabilities that far exceed those initial attempts, including the ability to generate discharge summaries and synthesize complex patient histories automatically. Currently, Care New England has deployed these tools to approximately 150 physicians across facilities including Kent Hospital, Women & Infants, and Butler Hospital.
The adoption is not limited to smaller systems. Brown University Health, the state’s largest health provider, has implemented similar AI tools for 500 primary care and emergency medicine physicians. Adam Landman, chief digital information officer at Brown Health and a practicing emergency doctor, notes that the goal is to shift the clinician’s cognitive load away from "typing furiously" and back toward listening and observation. Beyond the exam room, Brown Health is exploring AI for administrative utility, such as drafting grant proposals and managing operational problem-solving.
It is important to distinguish between the current application of these tools and the more speculative uses of AI in medicine. While headlines often focus on the potential for AI to autonomously diagnose patients or interpret high-stakes medical imaging, the current rollout in Rhode Island is strictly focused on administrative efficiency. Gregorio is explicit about this boundary, noting that Care New England intentionally avoids the "bleeding edge" of diagnostic AI. The priority remains ensuring that any tool deployed is "bulletproof" and "battle-ready" before it touches clinical decision-making.
Limitations to this transition exist, primarily regarding the delicate balance of trust between human judgment and algorithmic assistance. While the current cohort of doctors appears to be embracing the technology with high enthusiasm, the long-term impact on clinical accuracy and the potential for "automation bias"—where a clinician might over-rely on an AI-generated summary without proper verification—remains a subject for ongoing oversight.
The next stage of this evolution will be measured by the objective reduction in physician burnout rates and patient satisfaction metrics. As health systems continue to deploy these tools, the trajectory of clinical practice in Rhode Island will depend on whether these AI assistants can reliably sustain their performance without compromising the integrity of the medical record. For now, the move toward automating the clerk-work of medicine suggests a future where the doctor-patient relationship is no longer mediated by the relentless demands of a laptop.
